Speleothems As Environmental Recorders a Study of Holocene Speleothems and Their Growth Environments in Sweden

Speleothems as environmental recorders A study of Holocene speleothems and their growth environments in Sweden

Hanna S. Sundqvist

Thesis Contents

This doctoral thesis consists of a summary and four appended papers which will be referred to in the text by their Roman numerals.

List of papers

I Sundqvist, H. S., Seibert, J. and Holmgren, K. Understanding conditions behind speleothem formation in Korallgrottan, northwestern Sweden. Submitted to Journal of Hydrology.

II Sundqvist, H. S., Holmgren, K. and Lauritzen, S.-E., 2007. Stable isotope variations in stalagmites from northwestern Sweden document changes in temperature and vegetation, during early Holocene. The Holocene, 17(1).

III Sundqvist, H. S., Holmgren, K., Moberg, A., Spötl, C. and Mangini, A. Stable oxygen isotopes in a stalagmite from Jämtland, NW Sweden, record large temperature variations over the last 4000 years. Submitted to Geophysical Research Letters.

IV Sundqvist, H. S., Baker, A. and Holmgren, K. 2005. Luminescence in fast growing stalagmites from Uppsala, Sweden. Geograﬁska Annaler, 87 A (4): 539-548.

The co-authorship of Papers I-IV I have planned and designed the study, carried out the field work, performed most of the analyses and led all the paper writing. In paper I Jan Seibert did the HBV modelling. In paper II Stein Erik Lauritzen contributed with data from a Norwegian cave and the samples were dated at his laboratory. Christoph Spötl and Augusto Mangini contributed to paper III by being responsible for the stable isotope and dating laboratories used. Anders Moberg contributed to the scientific discussion and to the choice of statistical treatment of data in paper III. In paper IV, I and Andy Baker jointly performed the luminescence analysis. The thesis project was initiated by Karin Holmgren who made the first visits and did the first sampling in them. She has also participated several times in field and contributed to all papers through scientific discussions and improvements to the text.

1 Hanna S. Sundqvist

Introduction Project objectives

To understand human impact on climate today we need The overall aim with this PhD thesis is been to contrib- to reconstruct and understand how climate has varied ute highly resolved regional palaeoenvironmental infor- back in time, well before the industrial revolution. mation for the Holocene time period through studies of Knowledge about natural climate variability is also speleothems and their growth environments at different essential in studies of how human societies have been locations in Sweden. The specific objectives were: affected by and have adapted to climate change in the past. Climate change can be studied in a number of 1. To determine if speleothems from three Swedish natural archives. Speleothems, or cave drip stones are sites are suitable as palaeoclimatic archives. one of those archives which, like ice cores, tree rings, peat and lake sediments, can be used for reconstruct- 2. To obtain an understanding of factors that control ing past climate change. Carbonate speleothems are and affect the growth and properties of the spele- secondary deposits of calcium carbonate, chemically othems at these specific sites. precipitated in caves from carbonate seepage water. They are deposited either through degassing of car- 3. To examine if stable isotopes, luminescence and bon dioxide or through evaporation (Schwarcz 1986). trace elements in the speleothems can be used as Speleothems are well suited for uranium-series dat- proxy indicators of past environmental change. ing, producing ages directly in calendar years (Smart 1991, Ivanovich and Harmon 1992) and the mecha- 4. To provide palaeoclimatic data for the Holocene nisms controlling speleothem growth are sensitive to time period and compare it with other proxy data external, often climatically driven, processes. There- from the region. fore, many variables that can be measured in speleothems, such as stable carbon and oxygen isotopes, laminations and trace element composition, may serve as climatic proxies (Lauritzen and Lundberg 1999a). Speleothems as climatic archives Even though the advantages are acknowledged, problems still exist in using speleothems as palaeoclimatic Speleothems often hold a remarkable archive of data, archives (Smart and Richards 2003). Like other cli- describing local and global climatic and environmen- mate archives, speleothems are individuals. Each spe- tal conditions for the period of time in which they grew leothem is unique with its own response to external (Henderson 2006). The first palaeoclimatic studies of spe- processes; therefore ideally more than one sample leothems were conducted almost forty years ago (Hendy from each cave should be used. However, this ambi- and Wilson 1968, Thompson et al. 1974). The most im- tion can conflict with issues concerning cave conser- portant analytical development since the first studies is vation. Furthermore, karst processes are characterised the improvements in the technique of dating speleothems. by non-linear and threshold responses. To understand Unlike other continental archives speleothems are not these, laboratory and field experiments are needed to commonly dated with the radiocarbon method owing to determine the specific factors affecting speleothem the variable proportion of “dead” carbon from the bedrock deposition at different sites. (Genty et al. 2001). Instead speleothems are almost ideal In Sweden, caves with speleothems are found in for U-series dating. Uranium (238U, 235U, 234U) is bound Lummelundagrottan, developed in Silurian limestone, in silicate and oxide minerals in rocks and is co-precipi- on the island Gotland, and in caves developed in tated with the calcium carbonate forming the speleothem. Precambrian limestone or marble in the Caledonian It decays into daughter isotopes (230Th, 232Th), and as tho- mountain range in northern Sweden (Engh 1981). The rium is insoluble in water, is therefore absent at the time speleothems in Lummelundagrottan have been thor- of deposition. The production of 230Th from disintegra- oughly described by Engh (1981). A study of growth tion of 234U thus serves as a measure of time (Smart 1991, layers in a stalagmite from Lummelundagrottan was Ivanovich et al. 1992). However, contamination of the performed by Carlsson (1998). While the speleothems daughter isotope 230Th may occur as a result of co-pre- in Lummelundagrottan have proved to be difficult to cipitation of 232Th attached to clay particles present in the date, because of low uranium content in the Silurian cave drip water. However, it is possible to corrected for limestone, the speleothems from the mountain caves this. Another problem is that results from different dating contain enough uranium for precise uranium series labs are not strictly comparable because of the different dating, using available high resolution dating tech- spike calibration and correction methods (McDermott et niques. al. 2006). The dating techniques are straightforward and

2 Speleothems as palaeoenvironmental recorders

have been described thoroughly elsewhere (Smart 1991, Oxygen isotopes Ivanovich 1992). The advances in Thermal Ionisation The stable oxygen isotope composition of speleothem Mass Spectrometry (TIMS) have increased precision and calcite formed in isotopic equilibrium is related to the reduced the sample size, making it possible to achieve isotopic composition of the drip water and the cave chronological precision as good as 0.3-0.6 % (Lauritzen temperature. Studies have shown that the δ18O composi- and Lundberg 1999b). The recently introduced high-res- tion of cave drip water can be constant throughout the olution magnetic-sector field ICP-MS (inductively cou- year, and be approximately equal to the composition pled plasma mass spectrometry) techniques require even of the mean annual value of the outside precipitation smaller sample sizes than the TIMS-technique (Halliday (Schwarcz et al. 1976, Younge et al. 1985) and that the et al. 1998, Shen et al. 2002). In this thesis, the main deep cave temperature is close to the mean annual tem- speleothem variables studied are stable isotopes and lu- perature outside the cave (Wigley and Brown 1976). minescent laminations and therefore a short summary of The δ18O composition of precipitation is a consequence the theory behind these proxies is given. of several factors such as latitude, altitude, distance from sea, amount of precipitation and air temperature Stable isotopes reflecting the mass-fraction of moisture precipitated 18 When there is slow degassing of CO2, stable temperature from clouds. The mean annual δ O composition of pre- and no evaporation, calcium carbonate can be precipi- cipitation decreases systematically across Europe with tated in isotopic equilibrium with the parent drip water increasing distance from the North Atlantic Ocean (Ro- (Hendy 1971). This means that the partitioning of light zanski et al. 1993). Under present-day conditions, the and heavy isotopes between the aqueous and solid phase temperature dependence in rainfall is about 0.59 °± 0.09 is only a function of cave air temperature (Lauritzen and ‰°C-1 for European sites (Rozanski et al.1993). Since Lundberg 1999b). The calcite-water fractionation is – this exceeds the –0.24 ‰°C-1 calcite-water fractiona- 0.24 ‰ °C-1 (OŃeill et al. 1969). Characteristic features tion, a net shift to heavier O isotopes with higher tem- of equilibrium deposits are according to Hendy (1971): peratures is expected (i.e. a positive correlation between δ18O in the calcite and temperature)(Tab.1). Depending 1. Insignificant changes in stable oxygen isotopic com- on the site however, the temperature dependence in pre- position of calcite along a single growth layer. cipitation could be greater than, equal to, or less than the temperature dependence of δ18O in calcite deposited in 2. Any slight variation in stable oxygen isotopic com- speleothems (McDermott 2004, McDermott et al. 2006). position does not correspond to similar changes in the Also, the variable mixing of winter and summer pre- stable carbon isotopic composition of calcite along cipitation in the cave drip water could lead to a negative the same growth layer. instead of a positive relationship between δ18O in speleothems and temperature (e.g. Lauritzen and Lundberg This can be tested by analysing several calcite samples 1999b, Sundqvist et al. 2007, Vollweiler et al. 2006). along a single growth layer and is referred to as the Hendy On long timescales, factors other than temperature may test. Although it is a scientifically sound test in theory, it also cause temporal variations in δ18Op. These include; is in reality often impossible to apply, since the size of (i) changes in the δ18O of the oceanic source region of stalagmite growth layers normally are less than the size precipitation, (ii) changes in the temperature difference of the drill used (0.5-1 mm) for extracting the sample. It between the ocean surface temperature in the vapour has also been suggested that while non–equilibrium con- source area and the air temperature at the site of inter- ditions occur at the flank of the stalagmite, equilibrium est, (iii) shifts in moisture sources or storm tracks, (iv) conditions still may prevail in the centre of the stalagmite changes in the proportion of precipitation that has been (Dulinski 1990). Thus a negative result of the Hendy test derived from non oceanic sources, and (v) the amount of will not necessarily indicate that kinetic fractionation has precipitation (Gascoyne 1992). occurred. Isotope ratios in carbonates are expressed in Under favourable conditions, the δ18O signal of calcite the δ notation in parts per mille (‰) relative to V-PDB can be transformed into absolute temperature. This can (Vienna Pee Dee Belemnite), a calcite fossil from a lime- be achieved by measuring the isotopic composition in stone formation. Water samples are expressed in relation fluid inclusions in the speleothem (Harmon et al. 1979, to V-SMOW (Vienna Standard Mean Ocean Water): Gascoyne et al. 1981, Goede et al. 1986, Younge et al. 1981). Fluid inclusions, representing fossil seepage wa- δ18O V-PDB or SMOW = (Rsample/Rref – 1)*1000, ter, are formed in speleothems when small volumes of where R = 18O/16O drip water become trapped within the precipitating cal- Similarly, δ13C V-PDB = (Rsample/Rref – 1)*1000, cite (Rowe et al. 1998-1999). Methods for extracting and where R = 13C/12C analysing fluid inclusion are still under development.

3 Hanna S. Sundqvist

Carbon isotopes predominantly originates from organic acids trapped in- Less attention has been paid to the stable carbon iso- side the speleothem calcite. These acids occur in the over- tope composition of speleothem calcite. The δ13C can lying soil and are due to the breakdown of plant material be influenced by a number of different factors such as (Baker et al. 1993, Shopov et al. 1994). Organic acids are photosynthetic pathways (C3/C4 pathways), biologi- divided into two groups, fulvic and humic acids. Fulvic cal activity, overlying bedrock, rainfall, and drip rate acids are produced by photosynthesis and are released inside the cave (Baker et al. 1997, McDermott et al. through the plant roots. They are readily soluble and are 2006). when studying environments where only one or expected to enter speleothem feed waters preferentially two of these processes dominate a more exact interpre- during the growing seasons. Humic acids are products of tation of the signal is possible. Plants utilizing the C4 organic decomposition in the soil and the epikarst below, photosynthetic pathway respire and decompose into and dissolve more slowly. Studies of soil extracts suggest carbon dioxide with higher δ13C values than plants uti- that fulvic acids have a more intense luminescence and lizing the C3 pathway. C3 plants are more abundant in shorter wavelengths than the humic acids (Senesi et al. cool and moist climates whereas C4 plant predominate 1991). Together, the two groups can be taken as indices in warm and arid environments (Dorale et al. 1992). of productivity in the overlying soil and plant cover, and In high latitudes the vegetation uses only the C3 pho- therefore, as a proxy measure of palaeoclimate (Shopov tosynthetic pathways, and thus changes in δ13C must et al. 1994). However, climate is only one among many be explained by other mechanisms than fluctuations of factors, such as pH and metal-ion interactions, that may C3-C4 plants (Tab. 2). Changes in δ13C can be a result influence fluorescence intensity and the wavelength of of changes in biomass even in regions with only C3 dissolved organic matter in the soil and in natural waters vegetation (Baldini et al. 2005, Sundqvist et al. 2007). (McGarry and Baker 2000). In these regions decreases in the production rate of

CO2 as well as a decreasing vegetation density, give rise to higher δ13C values. Natural changes in vegeta- Field areas tion are often caused by changes in temperature and/or changes in humidity which also can cause changes in The ﬁeld work was carried out at three sites in Sweden δ13C (Linge et al. 2001, Niggeman et al. 2003, Onac (Fig. 1): the caves Korallgrottan (Papers I, II, III) and Lab- et al. 2002). Some temperate-zone speleothems show yrintgrottan (Paper II) in the Caledonian mountain chain values for δ13C that are greater than -6‰; these val- and a cellar vault in the city of Uppsala (Paper IV). ues are higher than would be predicted if they were in equilibrium with the expected C3 vegetation (Baker et al. 1997) and should therefore be interpreted with care. Such changes are likely to be caused by kinetic effects, IA like evaporation or rapid degassing of cave drip waters RUSS or calcite precipitation in the unsaturated zone above the cave. Ar ctic circle Luminescent laminae Labyrintgrottan Variations in crystal fabric, organic content, or trace con- Korallgrottan taminant content along the growth axis of the speleothem can be detected as various forms of laminae. These lami- D nae are either directly visible or can be produced by opti- FINLAN cal excitation of the sample. It has been shown that both NORWAY visible and luminescent laminae may be annually formed; however any timescale, from seasonal or single events to Uppsala centennial-millennial timescales and up, may be repre- A ESTONI sented by a lamina signal (Baker et al. 1993, Shopov et al. SWEDEN 1994, Genty et al. 1997, Tan et al. 2006). VIA Many speleothems exhibit luminescence (light emis- LAT DENMARK sion) when exposed to ultraviolet light sources. Absorp- NIA LITHUA tion of energy by atoms in the mineral leads to the rising of electrons from the ground energy state to an excited level. Later these electrons fall down to a lower level while emitting light (photons). Speleothem luminescence Figure 1. Map of Scandinavia with ﬁeld areas indicated.

4 Speleothems as palaeoenvironmental recorders

Korallgrottan (64°53’N, 14°9’E, 540-600 m a.s.l) in northern Jämtland is the largest known cave in Sweden, with approximately 5.5 km of passages and has been described previosly by Isacsson (1994). Although most of the cave is closed to the public, guided tours with groups of 10-15 people, in a restricted area, are arranged between June and October. Korallgrottan has developed in a 200 to 300 m wide belt of Precambrian limestone. The bedrock surface above the cave is overlain by a 30-50 cm thick soil cover. The vegetation above the cave is rich in herbs, grass and mosses but also consists of sparse forest with old spruce and birch (Fig. 2). Nearby meteorological stations have an mean annual precipitation of 866 mm Figure 2. View of the area above Korallgrottan. (Ankarvattnet 1961-90, 4 km south east) and an annual mean temperature of 1.4°C (Gäddede 1961-90, 50 km south) (Alexandersson and Eggersson Karlström 2001). From November to April most of the precipitation falls as snow and the snow cover generally lasts until May. Labyrintgrottan (66°3Ń, 14°41É, 730 m a.s.l) is more than 2.4 km long and has developed in a limestone probably Ordovician in age that is partly crystalline and folded (Helldén 1973). The area is part of the most westerly of the extensive limestone branches that stretches from the central parts of Artfjället down to the northernmost slope of Lake Överuman. The entrance of the cave is located about 50 m above the present day tree limit. The bedrock is in places covered by a thin (few cm thick) soil cover with herbs and grasses (Fig. 3). A nearby meteorological station (Hemavan 1961-90, 30 km southeast) has an annual mean precipitation of 748 mm and an annual mean temperature of -0.5°C (Alexandersson and Eggersson Karlström 2001). From October to late April most of the precipitation falls as snow and the snow cover generally remains until June. The city of Uppsala is situated in southeastern Sweden (59°54’N, 17°48’E) (Paper IV). Below the ground, next Figure 3. View of the area above Labyrintgrottan to Uppsala castle, two cellar vaults are situated (Fig. 4). The cellars were connected to a house that, together with parts of the castle, was destroyed in a fire in 1702. Both the cellar vaults and the house were probably constructed in the 1560s. These cellars, thought to have been completely demolished, where rediscovered by accident in 1976. To- day the cellars are situated beneath a 4-9 m thick layer of soil and are in this way closed to the public. Access to the cellars is possible every fifth year when inspection of the ground subsidence is made. Stalagmites and stalactites of considerable size have developed inside the cellars. These speleothems must have formed after the fire in 1702 when the house was destroyed and water began to be able to pen- etrate through the ground. The speleothems are most likely a result of the dissolution of the calcareous mortar between the bricks in the roof. Uppsala (1961-90) has an annual mean Figure 4. View of the area above the cellar vaults in precipitation of 545 mm and an annual average temperature Uppsala. of 5.6 °C (Alexandersson and Eggersson Karlström 2001).

5 Hanna S. Sundqvist

Methods drilled at intervals of 0.5-1 mm. Results from a study comparing both techniques (Spötl & Mattey 2006) show that All methods are described brieﬂy in this chapter. For the drilling data tend to miss extreme values picked up by more detailed descriptions, see papers in appendices I- the micromilling data. IV. In Paper II drilling was conducted using a standard dentist drill (0.5 mm in diameter). Vertical holes down to Fieldwork and monitoring about 2 mm in depth were manually drilled at 1 mm in- Stalagmites were sampled at the three sites. Stalagmites crements. The samples were analysed with Isotope Ratio K1 (Paper II) and K11 (Paper III) were sampled in Ko- Mass Spectrometry (IRMS) at the Department of Geology rallgrottan during visits in 1998 and 2005 respectively, and Geochemistry, Stockholm University. The IRMS in- L4 (Paper II) was sampled in Labyrintgrottan during a strument used is a Finnigan Mat 252, equipped with an on- visit in 1998 and U3 and U4 (Paper IV) were sampled line Kiel device. In Paper III micromilling was performed in the cellar vault in Uppsala in 2000. An automatic at 0.1 mm steps. A trench about 2 mm wide and 0.2 mm sampling station, registering drip rate, conductivity of deep was micromilled concordant to the lamination; these drip water, air pressure, air moisture, and temperature, are proportions suggested by Fairchild et al. (2006). Both was installed 200 m from the nearest entrance in Ko- the micromilling and the mass spectrometry analysis were rallgrottan in June 2000. Air temperatures were also performed at the Department of Geology and Paleontolo- measured at four other locations inside the cave, us- gy, the University of Innsbruck. Isotope ratios are reported ing small temperature sensors. Stalactite drip water was in the δ-notation relative to the V-PDB standard in parts ‰ collected during each visit for chemical analysis. The with a precision better than ± 0.1 ‰. pH and electrical conductivity of the drip water were measured in situ when collecting the stalactite drip water. Visits to Korallgrottan were made once or more every year between 2000 and 2006. From April 2005 to April 2006 the cave was visited every second month. Water samples, both from different locations within the cave, and from the precipitation outside Korallgrottan, have also been collected. Information on the isotopic composition of local precipitation was obtained from observations at Bredkälen (1975-88) about 200 km south of Korallgrottan (Burgman et al. 1980, Calles and Westman 1989).

Laboratory analysis All chemical analyses of stalactite drip water were performed at the Department of Geology and Geochemis- Figure 5. Conventional drilling try, Stockholm University (Paper I). The stalactite drip water was analysed for oxygen and hydrogen isotopes on a Finnigan Mat Delta Plus mass spectrometer, ani- - - - 2- ons (F , Cl , NO3 , SO4 ) were analyzed on an Dionex IC DX 300 ion chromatograph and cations (Ca2+, Mg2+, Ba2+, Sr2+, Na2+) on an Inductively coupled plasma Var- ian Vista AX. Saturation indices and pCO2 where calcu- lated using the Phreeqc Interactive programme. Samples for analyses of stable oxygen and carbon isotopes in speleothem calcite were sampled using two different techniques, manual drilling using a dentist drill (Paper II) (Fig. 5) and semi automated micromilling (Paper III) (Fig. 6). The main difference between the methods is that micromilling techniques are capa- ble of continuously sampling sections of speleothems (e.g. Frappier et al. 2002) at spatial resolution ranging from few tens of microns to fractions of a millimetre, Figure 6. Micromilling as compared to conventional drilling where samples are

6 Speleothems as palaeoenvironmental recorders

Luminescence analyses were performed on stalag- warm and dry summers, together with an early start of mites from Uppsala (Paper IV). A complete representa- the winter, can cause a hiatus in the speleothem depo- tion of the luminescence properties down the central axis sition at this site. The geochemical composition of drip of the stalagmites were collected, using a Perkin-Elmer water is fairly stable throughout the year, especially at LS-50B luminescence spectrophotometer with a Perkin- the slow-dripping sits. This is explained by the presence Elmer fibre-optic extension and moving stage, in order of storage and mixing of water in the bedrock above the to construct excitation-emission matrixes. Continuous cave gallery. The cave temperature at the monitoring site luminescence intensity records were obtained by scan- is close to the atmospheric annual mean temperature out- ning the stalagmites at fixed excitation and emission side the cave, while closer to the entrance temperatures wavelengths. To avoid problems with laser light scatter- show clear seasonal variations. Assuming that recent ing, which can cause the emitted signal to reflect calcite conditions can be applied back in time, the results from porosity, at least three scans were performed down and this study suggest that fossil speleothems selected from parallel to the central axis of the sample. slow-dripping sites deep inside the cave are suitable for The speleothems studied in this thesis were dated at palaeoclimatic reconstruction of variations at annual or the U-series dating laboratories at the University of Ber- slightly longer time-scales. gen, Norway, on a Finnigan MAT 262 mass spectrometer (Paper II) and at the University of Heidelberg, Germany Paper II on a Finnigan MAT 262 RPQ mass spectrometer (Paper Sundqvist, H. S., Holmgren, K. and Lauritzen, S.- III). E., 2007. Stable isotope variations in stalagmites The speleothem from Uppsala was dated with AMS from northwestern Sweden document climate and radiocarbon dating at the Ångström Laboratory at Upp- environmental changes during early Holocene, The sala University and calibrated with INTCAL 98 (Stuiver Holocene, 17(1). et al. 1998). This paper presents two early Holocene (9.6-5.9 ka BP) high-resolution stable isotope records of stalagmites from Results northwestern Sweden. The two stalagmites were collected from two caves, Labyrintgrottan and Korallgrottan, The following is a short summary of the four papers situated above and below today’s tree-limit, respectively. included in this thesis. The stable oxygen isotope records of the stalagmites con- firm previous observations of the temperature evolution Paper I during the early Holocene with a gradual warming from Sundqvist, H. S., Seibert, J. and Holmgren, K. Under- c. 9.6 ka BP, interrupted by cooler conditions at 8.5-8.0 standing conditions behind speleothem formation in ka BP. The results indicate that, albeit the close prox- Korallgrottan, northwestern Sweden. Submitted to imity to the North Atlantic, the cooler conditions were Journal of Hydrology. driven by two-to-three abrupt cold events rather than by one so-called 8.2 event only. The period between 10 to 8 The aim of this study was to investigate and character- ka BP seems to have been an instable period which was ise the environmental factors that control active spele- affected by the melting ice sheets. Except for the cold othem growth in Korallgrottan, northwestern Sweden events the stalagmite oxygen records show that between in order to assess whether fossil speleothems from the 7.8 to 5.9 ka BP temperatures in northwestern Sweden this site are suitable as palaeoclimatic archives. The cave were warmer than today, with the interval between 7.8 to microclimate and the δ18O signal in speleothems, cave 5.9 ka BP most likely being the warmest. drip water and precipitation were monitored periodically It is proposed that the high-amplitude changes in the between December 2000 and April 2006. The results stable carbon isotope record of Labyrintgrottan are pro- show that the drip rate in Korallgrottan varies substan- posed to reflect changes in local vegetation, whereby at tially following the seasons at fast dripping sites. Drip the time of stalagmite growth, between 9.5-7.5 ka BP the rates are highest during autumn and lowest during winter area above Labyrintgrottan was covered by much denser and very dry summers. At slow dripping sites however, vegetation than today; it is not unlikely that between 9 drip rate and chemical composition are almost constant and 8 ka the area was below the local tree-limit. Since throughout the year. The drip water reaches the highest temperatures are the dominant factor governing vegeta- saturation level during the summer and autumn when bi- tion in northern Scandinavia the stable carbon isotope ological activity is most intense and the partial pressure record from Labyrintgrottan support this interpretation of carbon dioxide, which control limestone dissolution, that the conditions in Early Holocene were warmer than is highest. A shortening of this period, such as during today, resulting in a denser vegetation cover, which could

7 Hanna S. Sundqvist

produce enough CO2 to support speleothem growth in stone caves. They are deposited from hyperalkaline wa- this high-latitude, high-altitude region. ters (pH>10) by the reaction Ca(OH)2 + CO2 → CaCO3

+ H2O which is termed “concrete carbonation”. Spele- Paper III othems deposited by concrete carbonation usually have accelerated growth rates, growing 1-20 mm a year. The Sundqvist, H. S., Holmgren, K., Moberg, A., Spötl, stalagmites in the Uppsala cellar vaults display laminae C. and Mangini, A. Stable oxygen isotopes in a sta- that are suggested to be annual. The stalagmites will thus lagmite from Jämtland, NW Sweden, record large represent a period of 10-15 years with growth rates of temperature variations over the last 4000 years. Sub- 3-8 mm per year, which is similar to other fast-growing mitted to Geophysical Research Letters. speleothems. Because the highest amounts of precipitation in Uppsala are received during the summer and au- Paper III discusses the result from high-resolution stable tumn it is likely that the luminescent lamina are formed isotope analyses of a stalagmite from Korallgrottan that sometime during the autumn when precipitation is high has been growing over the last 4000 years. The stable and temperatures relatively low. AMS radiocarbon dat- oxygen isotope signal is proposed to record changes in ing revealed ages that were older than the cellars them- the relative contribution of winter versus summer pre- selves indicating that at least in part, the carbon has been cipitation into the cave as a result of changes in seasonal deposited by another process than concrete carbonation. temperatures. Cold temperatures would lead to relatively Owing to the uncertainty of the age of these speleothems stronger influence of summer precipitation on the stable it was not possible to compare the results with available oxygen composition of the stalagmite resulting in en- meteorological data. In addition, the short amount of riched values, while warmer temperatures imply rela- time covered by stalagmite growth further limits the pos- tively more winter precipitation which would result in sibility of undertaking such analyses. depleted δ18O values in the speleothem. The temperature decrease, inferred from the δ18O record, is interrupted by a number of distinct warm and cold spells of a few hun- Discussion dred years in length. The stalagmite δ18O record agrees with the concept of a warm period, the so-called Medi- This thesis is the first detailed study of speleothems eval Warm Period (MWP), centred around AD 1000 and from Swedish caves for paleoenvironmental reconstruc- a cold period, the so-called Little Ice Age (LIA), some- tion. The monitoring study emphasizes the suitability where between AD 1000 and today. However, based of speleothems from Korallgrottan as environmental on the stalagmite results, it seems as if the minimum recorders. In addition, the similarities between contem- temperatures during LIA arrived 100-150 years earlier porary samples from Labyrintgrottan and Korallgrottan in northern Jämtland than in the northern hemisphere in demonstrate that also speleothems from Labyrintgrot- general. Closer-spaced precise age determinations of this tan are good proxies for the regional environment and stalagmite may yield new and detailed information on climate. Taken together, the stable isotope records (K1 the evolution of climate in northern Sweden. and K11) from Korallgrottan together cover much of the early and late Holocene, although there is a gap of Paper IV about 2000 years in the middle between 4000 and 6000 Sundqvist, H. S., Baker, A. and Holmgren, K. 2005. Lu- years BP (Fig. 7). When comparing the both records it minecence in fast growing stalagmites from Uppsala, appears that the late Holocene could have been more Sweden. Geografiska Annaler, 87 A (4): 539-548. variable than the early Holocene. However, this could In this paper the results of a study of luminescent prop- however be explained by the fact that the different sam- erties in fast growing stalagmites from a cellar vault pling techniques were the conventional drilling tend to in Uppsala, southeastern Sweden, are discussed. The miss extreme values beacause of the lower sampling hypothesis was that the stalagmites have been grow- resolution and discontinues sampling (Spötl and Mat- ing continuously during the past 300 years and that the tey 2006). The mean values of both of the δ18O records variations in luminescence and growth rate could be cali- vary around 9.1 ‰. It has however been reported that brated against available meteorological data. The results, speleothems from the same cave can have δ18O records however, indicate that the stalagmites have been grow- varying around different mean values still displaying ing during much shorter periods of time. These periods the same pattern of variability (Linge et al. 2001, Voll- have been difficult to determine exactly, using available weiler et al. 2006). Since the records of K1 and K11 do dating methods. The stalagmites consist of calcium car- not overlap it is hard to say if this is the case for K1 and bonate but are precipitated by a different process than K11. Thus it is not possible to judge whether the early natural calcium carbonate speleothems found in lime- Holocene was warmer or colder than the late Holocene.

8 Speleothems as palaeoenvironmental recorders

Calendar years representability of stalagmite δ18O records as recorders 6000 4000 2000 BC/AD 2000 of temperature changes, the δ18O record of stalagmite -10.0 -10.0 K11 from Korallgrottan (Paper III) is compared here to previously published records of stalagmite δ18O records -9.5 -9.5 from northern and central Europe (Fig. 8, Tab. 3). The records in the comparison were chosen beacause of due VPDB) -9.0 -9.0 to their relative geographical closeness to each other

O (‰ and their influence by North Atlantic climate. Two of 18 δ -8.5 -8.5 the records, SG93 from northwestern Norway (Lau- K1 K11 ritzen and Lundberg 1999b, Linge et al. 2001) and COMNISPA from Austria (Vollweiler et al. 2006), both -8.0 -8.0 8000 6000 4000 2000 0 have been interpreted as showing a negative correla- 18 Years BP tion between speleothem δ O and past temperatures, mainly because the speleothems are believed to receive a higher proportion of light winter precipitation during Figure 7. Both δ18O records from Korallgrottan (K1-K11) low warm winters. COMNISPA is a combination of three pass filtered (gaussian) corresponding roughly to a 100-year stalagmite δ18O records from Spannagel cave in Austria running mean. that together cover the last 9000 years. In this comparison which covers the last 4000 years COMNISPA is Thus the results in this thesis contribute thus new infor- composed of the record of stalagmite SPA 12, which mation about the suitability of using small stalagmites spans between 70 and 2560 years ago and the average from caves in the Caledonian mountains for high resolu- curve of stalagmites SPA 12 and SPA 128 from between tion reconstruction of relative changes in regional tem- 2560 and 4000 years ago. In the third stalagmite record, perature, humidity and vegetation. The new techniques stalagmite CC3 from southwestern Ireland (McDermott recently developed for stalagmite analyses provide new et al. 1999, 2001), the δ18O record is believed to have possibilities for sampling at high resolution. This implies a positive correlation with temperature. The strongest that continued analyses of stalagmites from the caves argument for this hypothesis is the positive correlation studied for this thesis can contribute information on the between δ18O and growth rate. The comparison reveals evolution of the Holocene climate in northern Scandina- both similarities and differences between the records. via at a time resolution that is of yet largely unavailable, Differences in the records that are easily explained except for in tree ring records (Briffa et al. 1992, Grudd are, for example, the variation in amplitudes between et al. 2002, Gunnarson et al. 2003). The interpretation the records (Tab. 3), with the Irish record having the of δ18O composition in stalagmites is however still not highest amplitude of 2.0 ‰ compared to the Norwe- straight-forward. Since past variation in the δ18O content gian record, which has the lowest amplitude of 0.8 ‰. of drip water is unknown at the sites studied it is not pos- This difference can partly be explained by differences sible to retrieve absolute temperature variations. It is also in bedrock thickness, with the Norwegian cave having not possible to objectively determine exactly the propor- a bedrock thickness of about 100 m which results in tions of local site-specific influences versus regional more mixing of waters of different ages, as compared to changes. The solutions to these problems rely much on the others having thicknesses in the order of 10-20 m. a knowledge of regional-local climate and atmospheric In like with the precipitation patterns in Europe today circulation conditions as well as on on comparison with (Rozanski et al. 1993) the mean value of the oxygen other data. In paper II and III the stalagmite results were isotopes decrease with distance from the Atlantic ocean interpreted and discussed in relation to a number of dif- and with latitude/altitude. This explains why the Swed- ferent climate series. The following section discusses ish record displays the lowest isotope values while the comparison with additional data that were not included Irish displays the highest. Even if the mean values and in the appended papers. amplitude differences between the records can be explained it is obvious that many differences still exists. 18 Comparison with other δ O speleothem records from Similarities in the records of the last millennium in- northern and central Europe clude the Little Ice Age drop in temperature apparent at The δ18O signal of the analysed stalagmites (Papers II AD1300-1800 in K11, at AD1400-1800 in COMNISPA and III) is inferred to be of regional character and pri- (SPA 12) and the two two-stage drop at 1400-1800 in marily reflect changes in temperature. Similarities be- CC3. In SG93 it is expressed as a general decline from tween the stalagmite records and other climate proxy AD 1200 until recent times. A Medieval Warming is records support this hypothesis. As a test of the regional apparent at AD 800-1200 in CC3, AD 1100-1300 in

9 Hanna S. Sundqvist

2000 1500 1000 500 BC/AD 500 1000 1500 2000 -2

a) -3 VPDB)

-4 3 (‰ O CC 18 δ -5

VPDB) -8 b)

-7 O SG93 (‰

18 -10 δ c) VPDB) -9 VPDB)

-9 O K11 (‰ 18 -8 δ

A (‰ d) -8 OMNISP O C 18

δ -7 2000 1500 1000 500 BC/AD 500 1000 1500 2000

Figure 8. A comparison between four different δ18O stalagmite records from Western Europe. a) CC3, Crag Cave, Ireland (McDer- mott et al. 1999, 2001), b) SG93, Søylegrotta, Norway (Lauritzen and Lundberg 1999b, Linge et al. 2001), c) K11, Korallgrottan, Sweden, this thesis, d) COMNISPA, Spannagel cave, Austria (Vollweiler et al. 2006). Note that the δ18O scales are reversed for SG93, K11 and COMNISPA. The records have been low pass ﬁltered (gaussian) corresponding roughly to a 100-year running mean.

SG93, AD 900-1100 in K11 and COMNISPA (SPA12). factor inﬂuencing the oxygen isotope signal of these sta- Both the record from Ireland and the one from Sweden lagmites. The amount of rainfall, or other local factors, have slightly increasing δ18O values during the last 4000 also could have controlled the oxygen isotope composi- years, which is interpreted to reﬂect a gradual temperature tion also. Further tests and comparisons with other proxy decrease at the Swedish site and as a temperature increase records can tell which of these records are the best proxys at the Irish site. A possible explanation for this could be of past temperature changes. that the increase observed in both stalagmite records is caused by changes in the atmospheric circulation pattern Comparison between stable isotope data of K11 and a or in the isotopic composition of the source water. tree-ring chronology from Jämtland It is evident that interpretation of temperatures made One of the conclusions from Paper III is that the δ18O from δ18O variations in speleothems is complicated and record of stalagmite K11 captures regional temperature as can be seen by the differences existing between the variations. The δ13C record has some features similar four records it is also clear that temperature is not the only to the δ18O record, which indicates that δ13C is partly

10 Speleothems as palaeoenvironmental recorders

2000 1500 1000 500 BC/AD 500 1000 1500 2000 2.0 (a)

x 1.5 warmer

1.0 ee-ring inde

-10.0 0.5 (b) -9.5

V-PDB 0.0

O ‰ -9.0 -10 18 δ (c) -9 -8.5

-8 V-PDB C ‰

-7 13 δ

-6 (d) 0 -5 10

lapping 20 higher lake lavels

Number of trees over 30 (wetter)

2000 1500 1000 500 BC/AD 500 1000 1500 2000

Calendar years

Figure 9. A comparison between the Håckren tree-ring chronology (SW Jämtland) (Gunnarson et al. 2003) and stable isotope data from K11, Korallgrottan (NW Jämtland). (a) tree-ring index from Håckren (b) stable oxygen isotopes of K11 (c) stable carbon isotopes of K11 (d) sample depth changes through time of the Håckren chronology. Data in (a)-(c) are low pass ﬁltered (gaussian) corresponding roughly to a 100-year running mean.

11 Hanna S. Sundqvist

is affected by temperature as well. It is proposed that Conclusions the δ13C signal reflects changes in the vegetation density, which in turn is governed by a combination of With careful analysis and interpretation, it seems pos- temperature and humidity changes. The high-ampli- sible to use stable isotopes in fossil speleothems from tude changes in the stable carbon isotope record of Labyrintgrottan and Korallgrottan to provide high reso- Labyrintgrottan were also proposed to reflect changes lution information on regional climate variations in in the local vegetation (Paper II). In order to test this northwestern Sweden. Oxygen isotopes can yield infor- hypothesis further, I here compare the stable isotope mation on relative changes in past temperature, where- records of K11 with the Håckren tree-ring chronol- as and carbon isotopes can provide information about ogy from SW Jämtland (Gunnarson et al. 2003) (Fig. relative changes in vegetation density. Still uncertain- 9). This data set was chosen for three reasons 1) it is ties regarding age precision and isotope interpretation geographically close to Korallgrottan (200 km south), exist, but much of these are likely to be overcome by 2) it is a perfectly dated calendar year record and 3) more precise dating and further analysis both regard- the tree-ring index (tree-ring width) capture variations ing general process studies and detailed isotope studies. in July temperature and the number of trees (sample In summary, the most important conclusions from this depth) is proposed to capture fluctuating lake levels study are: and thereby changes in humidity (Gunnarson et al. 2003). If the trees reflect temperature and humidity • The monitoring study in Korallgrottan underscores and the stable isotopes reflect temperature and vegeta- the importance of examining the local climatology, tion then low oxygen isotope values could be expected drip hydrology and geochemistry of drip water be- to correlate with high tree-ring index (discussed in Pa- fore selecting stalagmites for palaeoclimate analy- per III) and low values of the carbon isotopes to cor- sis. relate with high tree-ring index combined with high lake levels (few overlapping trees). High δ13C values • The drip rate from fast-dripping stalactites usually should then correspond to cold and/or very wet/dry shows larger seasonal and year-to-year variations periods. This is an initial comparison and many sourc- and reflects the situation of the snow regime while es of error exist, such as the limited number of datings slow- dripping ones have more stable drip rates. The on stalagmite K11 and the fact that the stable isotope slow- dripping stalactites also have a more stable, data comes from one stalagmite only. The ring-index probably annual, geochemical signal in the drip wa- curve and the δ18O curve, both of which are believed ter. to reflect temperature changes, have many similar features. The ring-width curve shows periods with en- • Stalagmites fed by stalactites with slow and stable hanced growth which coincide with low δ18O values drip rates from deep inside the cave are the most at 1400-1300 BC, 1000 BC, 500 BC, AD 900-1000 suitable as palaeoclimate archives. AD and AD 2000. The records do not match at 300- 100 BC and at AD 400-500. In general low δ13C val- • The close similarities between the δ18O records of ues correspond to high summer temperatures and high stalagmites from Labyrintgrottan and Korallgrottan lake level stands with one exception at 1300-1100 BC. emphasize the potential of the speleothems from High δ13C values correspond to low temperatures and both caves for providing high resolution regional low lake levels at 300-100 BC, AD 100-200 and AD palaeoclimate information. 1400-1600; high δ13C values correlate with high temperatures and high lake levels at 1600-1400 BC and • The δ18O signal in the stalagmites is interpreted as AD 900-1100. After a first attempt to compare these varying as a result of the variable mixing of sum- two records it seems that that carbon isotopes reflect mer and winter precipitation owing to changes in re- vegetation density or soil zone conditions might hold gional temperatures. The δ13C signal is interpreted as true for this site; however, a more accurate age model reflecting changes in soil CO2 production, which in and at least one more stalagmite from this cave would turn may reflect changes in vegetation density. be needed to validate these results. In addition, tree- ring indices contemporary with low sample depth, • The stalagmite δ18O records from northern Scandi- like at 500 BC, AD 500 and AD 900 should be treated navia indicate that between 9500 and 6000 years ago with caution since internal factors (e.g. microclimate, temperatures in northwestern Sweden were warmer diseases) unique to a single tree may have a large im- than today, except for a number of cold events. Dur- pact when chronologies consist of only a few samples ing this period the interval between 7800 and 6000 (Gunnarson et al. 2003). years ago seems to have been the warmest.

12 Speleothems as palaeoenvironmental recorders

• The area above Labyrintgrottan was most likely of about 3000 years and is already sampled for stable covered by much denser vegetation than today at isotopes for comparison and evolution of the results in the time of stalagmite growth (9500-7500 years ago) Paper III. and was - unlike today – probably situated below the Dating the base level of collected and not yet dated local tree-limit between 9000 and 8000 years ago. speleothems from Labyrintgrottan might reveal information about the time of deglaciation of the area. • The stable isotope record together with changes in A flowstone of Eemian age (125 ka) has been found growth rate of a stalagmite covering the last 4000 in Korallgrottan. The sample has been dated by alfa years indicate a slight but general reduction of tem- spectrometry but a new age model together with stable perature and precipitation during the growth pe- isotope analysis might yield valuable information about riod. The temperature decrease is interrupted by a the Eemian environment. It would also be interesting to number of distinct warm and cold spells of a few test whether the negative correlation between the oxy- hundred years in length. The record agrees with the gen isotopes and temperature holds this far back in time concept of a warmer period, the so-called Medieval or if other processes are more dominant. Warm Period (MWP), centred around AD 1000 and a colder period, the so-called Little Ice Age (LIA), Uppsala somewhere between AD 1000 and today. The next opening of the cellar vaults in Uppsala is planned to take place in 2007. During the last visit in • Fast growing speleothems, like the ones from Upp- 2000 plastic slides were placed out on the floor. Hope- sala, can provide an opportunity of studying lumi- fully calcite has precipitated on top of these. Then lu- nescence properties in speleothems in greater detail minescence properties can be studied with great age than otherwise possible. The results indicate that control and directly compared to meteorological data. the variations in luminescence intensity are annual and that the annual laminae of the luminescent Monitoring and modelling record represent a flush of organic material. With To examine the contribution of winter precipitation ver- good age control they can yield information about sus summer precipitation of the cave drip water, the an- history and climate. nual δ18O in the precipitation should be monitored and compared it with the annual δ18O in the drip water of a number of stalagmites (especially K11). This is a study Future perspectives that requires several years of sampling and a faster ap- proach would be to calculate the contribution of winter Carbon isotopes precipitation compared with summer precipitation using the HBV model. To further examine the hypothesis that carbon isotopes represent a vegetation density signal one could compare carbon isotopes with pollen data from a lake in Acknowledgements the vicinity of the cave. This type of study would be of particular interest in the Labyrintgrottan area. Pol- I have received financial support from Ahlmann, KVA, len data obtained from the lake sediments could pro- Lagrelius, Mannerfeldt, SSAG, Donationsstipendier, vide information about changes in tree-limit positions VR to Karin Holmgren. Länsstyrelsen Jämtland and that could possibly be coupled to the changes in δ13C Västerbottens län for permission for sampling the spe- in the Labyrintgrottan stalagmite. The lake should then leothems. be situated at the same elevation or slightly higher than This thesis would not have been finished if it was not for the cave. In addition, comparisons between carbon iso- a number of people helping me. First of all I would like topes and trace elements ratios such as Mg/Ca and Sr/ to thank my main supervisor Karin Holmgren for believ- Ca which together could yield information about dry ing in me and giving me the opportunity to do research as periods would be interesting. a PhD student. Thank you for letting me grow in my own time. Your optimism and enthusiasm is always there even Dating when I hand you the same half-written manuscript for the Additional palaeoenvironmental information from Ko- third time. My co-supervisors Regine Hock and Anders rallgrottan and Labyrintgrottan could be attained by Moberg have given valuable critical input on my man- dating more material. A better age model of K11 to- uscripts which have improved them a lot. I would also gether with an age model of stalagmite K13, sampled like to thank my co-writers: Andy Baker, Stein-Erik Lau- about 1 m away from K11, which has got a bottom date ritzen, Augusto Mangini, Jan Seibert, Christoph Spötl. I

13 Hanna S. Sundqvist

would also like to thank Magnus Mörth and Klara Haynal Baker, A., Ito, E., Smart, P. L. and McEvan, R. F., 1997. El- for helping me with the geochemical analyses and Steffen evated and variable values of 13C in speleothems in a British Holzkämper for helping me get access to various labs in cave system. Chemical Geology 136, 263-270. a time of panic.I have also received a lot help and interest Baker, A., Smart, P. L., Edwards, R. L. and Richards, D. A., from members of the Swedish Speleological Society espe- 1993. Annual growth banding in a cave stalagmite. Nature cially Mikaela Löfgren, Karl Grönvik, Christopher Krafft, 364, 518-520. Rolf Engh and Christer Örtwall. Baldini, J. U. L., McDermott, F., Baker, A., Baldini, L. M., During my 16 field trips to Korallgrottan I have had Mattey, D. P. and Railsback, L. B., 2005. Biomass effects about the same number of field assistants. You did not on stalagmite growth and isotope ratios: A 20th century ana- only crawl and drag the equipment in a muddy cave but logue from Wiltshire, England. Earth and Planetary Science also drove me up there (I did not have a drivers licence Letters 240, 486-494 until 2005). Thank you: Lisen Baier, Hanna Hansson, Ka- Briffa, K. R., Bartholin, T. S., Eckstein, D., Schweingruber, F. rin Holmgren, Martin Jonsson, Julia Lee Thorpe, Mikaela H., and Zetterberg, P. and Eronen, M., 1992. Fennoscandian Löfgren, Johan Mattsson, Anja Midttun, Oliver O’Hehir, summers from AD 500: temperature changes on short and Lena Rubensdotter, Ludvig Ruderstad, Oscar Ruderstad, long timescale. Climate Dynamics 7, 111-119. Pernilla Schuber, Madeleine Simon, Jonas Sundqvist, Burgman, J. O., Eriksson, E., Kostov, L. and Sundblad, G. Sara Sundqvist and Peter Westman. Special thanks to Oxygen-18 variation in monthly precipitation over Sweden. Oscar for handling the “Snow shoe case” against Rusta. Uppsala University, Department of Physical Geography, Di- Thomas Schneider have for several field trips lent me his vision of Hydrology, 24 pp. expensive pH meter, thank you for that. Calles, B. and Westman, F. 1989. Oxygen-18 and Deuterium in My heroes in Ankarvattnet, Bojan och Birger Edfors, precipitation in Sweden. Uppsala University, Department of thank you for your great hospitality. During my field trips Physical Geography, Division of Hydrology, Report Series to Korallgrottan I have stayed in you basement and you A No 47, 20 pp. have supplied me with precipitation data and samples, in- Carlsson, A., 1998. Analys av tillväxtlager i stalagmit från teresting weather conversations and a lot more. This thesis Lummelunda, Gotland. Examensarbete i naturgeografi, would probably look a lot different if it wasn’t for you. Department of Physical Geography, Stockholm Univer- Thank you also Ola Sundqvist from Rid I Jorm and Håkan sity, 41 pp. Berglund from Länstyrelsen in Jämtlands län. For those Dorale, J. L., Gonzales, L. A., Reagan, M. K., Pickett, D. A., of you who yet haven’t visited northwestern Jämtland. Go Murrell, M. T. and Baker, R. G., 1992. A high-resolution there! It is one of the most beautiful places in the world. record of Holocene Climate Change in Speleothem Cal- During the last month I have received a lot of help cite from Cold Water Cave, Northeast Iowa. Science 258, from friends and colleagues with reading manuscripts, 1626-1630. layout and other practical things. Thank you: Hanna H, Dulinski, M., 1990. δ18O in speleothems and its palaeocli- Christina, Henriette, Håkan, Björne, Lena, Tobbe and matic significance – some remarks. Freiberger-Forschung- Stefan. shefte,-Reihe-C 442, 72-83. Thank you to all fellow PhD-students and colleagues at Engh, L., 1981. Lummelundagrottan med tillhörande karstom- the Department of Physical Geography and Quaternary råde. Doctor scientiarum thesis, Lund University, 290 pp. Geology especially “lunchgruppen”, in all your different Fairchild, I. J., Smith, C. L., Baker, A., Fuller, L., Spötl, C., constellations, for being sort of a spiritual sanctuary for Mattey, McDermott, F., E. I. M. F., 2006. Modifaction the soul here at the department. and preservation of environmental signals in speleothems. Many thanks to my parents and parents in law for your Earth-Science Reviews 75 (1-4), 105-153. support and baby sitting Frappier, A., Sahagian, D., González, L. A. and Carpenter, S. Tack Martin, min superhjälte, för ditt engagemang, J., 2002. El Niño Events Recorded by Stalagmite Carbon stöd och kärlek som har hjälpt mig att få den här avhan- Isotopes. Science 298, 565. dlingen klar. Jag hade aldrig klarat det utan dig! Till Alba Gascoyne, M., 1992. Palaoeclimate determination from cave och Alfons: mamma har skrivit klart boken nu och kom- calcite deposits. Quaternary Science Reviews 11, 609- mer vara hemma och leka lite mera med er. 632. Gascoyne, M., Ford, D. C. and Schwarcz, H. P., 1981. Late References Pleistocene chronology and palaeoclimate of Vancouver Island determined from cave deposits. Canadian Journal Alexandersson, H. and Eggertsson Karlström, C., 2001. of Earth Science 18, 1643-1652. Temperaturen och nederbörden i Sverige 1961-1990. Ref- Genty, D., Baker, A. and Barnes, W., 1997. Comparison of erensnormaler – utgåva 2. Swedish Meteorological and Hy- annual luminescent and visible laminae in stalagmites. drological Institute, Meteorologi 99, Norrköping, Sweden. Earth and Planetary Science 325, 193-200.

14 Speleothems as palaeoenvironmental recorders

Genty, D., Baker, A., Maussalt, M., Proctor, C., Gilmour, M., from a δ18O stalagmite record. Earth and Planetary Sci- Pons-Banchu, E. and Hamelin, B., 2001. Dead carbon in sta- ence Letters 235, 741-751. lagmites: Carbonate bedrock paleodissolution vs. ageing of McDermott, F., 2004. Palaeo-climate reconstruction from sta- soil organic matter. Implications for 13C variations in spele- ble isotope variations in speleothems: a review. Quaternary othems. Geochimica Cosmochimica Acta 65. 3443-3457. Science Reviews 23, 901-918. Goede, A., Green, D. and Harmon, R. S., 1986. Late Pleistocene McDermott, F., Frisia, S., Huang, Y., Longinelli, A., Spiro, palaeotemperature record from a Tasmanian speleothem. B., Heaton, T. H. E., Hawkesworth, C.,Borasato, A., Kep- Australian Journal of Earth Sciences 33, 333-342. pens, E., Fairchild, I. J., van der Borg, K., Verheyden, S. Gunnarson, B. E., Borgmark, A. and Wastegård, S., 2003. and Selmo, E., 1999. Holocene climate variability in Eu- Holocene humidity fluctuations in Sweden inferred from rope: Evidence from δ18O, textural and extension rate dendrochronology and peat stratigraphy. Boreas 32, 347- variations in three speleothems. Quaternary Science Re- 360. views 18, 1021-1038. Grudd, H., Briffa, K. R., Karlén, W., Bartholin, T. S., Jones, P. McDermott, F., Mattey, D. P, and Hawkesworth, C., 2001. D. and Kromer, B., 2002. A 7400-year tree-ring chronology Centennial-Scale Holocene Climate Variability Revealed in northern Swedish Lapland – Natural climate variability by a High-Resolution Speleothem δ18O Record from SW expressed on annual to millennial time scale. The Holocene Ireland. Science 294, 1328-1331. 12, 657-665. McDermott, F., Schwarcz, H. and Rowe, P.J, 2006. Isotopes Halliday A.N., Lee, D.C., Christensen, J.N., Rehkämper, M., in speleothems. In. Leng, M. J. (Ed.), Isotopes in Pal- Yi, W., Luo, X., Hall, C.M., Ballentine, C.J., Pettke, T., and aeoenvironmental Research. Springer, Dordrecht, The Stirling, C., 1998. Applications of multiple collector-ICPMS Netherlands, 185-226. to cosmochemistry, geochemistry and paleoceanography. McGarry, S. F. and Baker, A., 2000. Organic fluorescence: Geochimica Cosmochimica Acta 62, 919-940. applications to speleothem palaeoenvironmental recon- Harmon, R. S., Schwarcz, H. P. and OŃeil, J. R., 1979. D/H struction. Quaternary Science Reviews 19, 1087-1101. ratios in speleothem fluid inclusions: a guide to variations Niggemann, S., Mangini, A., Richter, K. D. and Wurth, G. in the isotopic composition of meteoric precipitation? Earth 2003. A paleoclimate record of the last 17,600 years in and Planetary Science Letters 42, 254-266. stalagmites from the B7 cave, Sauerland, Germany. Qua- Henderson, G. M., 2006. Caving In to New Chronologies. Sci- ternary Science Reviews 22, 555-567. ence 313, 620-622. Onac, B. P., Constantn, S., Lundberg, J. and Lauritzen, S. Hendy, C. H., 1971. The isotopic geochemistry of speleothems E. 2002. Isotopic climate record in Holocene stalagmite -I. The calculation of the effects of different modes of forma- from Ursilor Cave (Romania). Journal of Quaternary tion on the isotopic composition of speleothems and their Science 17, 319-327. applicability as palaeoclimatic indicators. Geochemica et Rozanski, K., 1993. Isotopic patterns in modern global Cosmochica Acta 35, 801-824. precipitation. In: Climate Change in Continental Iso- Hendy, C. H. and Wilson, A. T., 1968. Palaeoclimatic data from topic Records, Swart, P. K., Lohmann, K. C., Mckenzie, speleothem. Nature 216, 4851. J. Savin, S. (eds). Geophysical Monograph 78, American Isacsson, G., 1994. Vad kan man se i Korallgrottan? Grottan Geophysical union: Washington, DC, 1-36. 2, 21-23. Schwarcz, H. P., Harmon, R. S., Thompson, P. and Ford, Ivanovich, M. and Harmon, R. S., editors, 1992. Uranium se- D. C., 1976. Stable isotope studies of fluid inclusions in ries disequilibrium. Applications to environmental problems. speleothems and their palaeoclimatic effect. Geochemica Clarendon, Oxford, 571 pp. et Cosmochimica Acta 40, 657-665. Lauritzen, S. E. and Lundberg, J., 1999a. Speleothems and Senesi, N., Miano, T. M., Provenzano, M. R. and Brunetti, climate: a special issue of The Holocene. The Holocene 9, G., 1991. Characterisation, differentiation and classifica- 643-647. tion of humic substances by flourescence spectroscopy. Lauritzen, S. E. and Lundberg, J., 1999b. Calibration of the Soil Science 152, 259271. speleothem delta function: an absolute temperature record Shen, C.C., Edwards, R.L., Cheng, H., Dorale, J.A., Tho- for the Holocene in northern Norway. The Holocene 9, 659- mas, R.B., Moran, S.B., Weinstein and S.E., Edmonds, 669. H.N., 2002. Uranium and thorium isotopic and concen- Linge, H., Lauritzen, S. E., Lundberg, J. and Berstad, I. M., tration measurements by magnetic sector inductively 2001. Stable isotope stratigraphy of Holocene speleothems: coupled plasma mass spectrometry. Chemical Geology examples from a cave system in Rana, northern Norway. 185, 165-178. Palaeogeography, Palaeoclimaatology, Palaeoecology Shopov, Y. Y., Ford, D. C. and Schwarcz, H. P., 1994. Lu- 167, 209-224. minescent microbanding in speleothems: high resolution Mangini, A., Spötl, C. and Verdes, P., 2005. Reconstruction chronology and palaeoclimate. Geology 22, 407-410. of temperature in the Central Alps during the past 2000 yr Schwarcz, H. P., 1986. Geochronology and isotopic geo-

15 Hanna S. Sundqvist

chemistry of speleothems. In: Fritz, P. and Fontes, J. Ch., (eds), Handbook of environmental isotope geochemistry. Vol 2. The terrestrial environment B, Elsevier, 271-303. Smart, P. L. 1991. Uranium series dating. In Smart, P. L. and Frances, P. D. (eds.) Quaternary Dating Methods: a users guide. Technichal Guide No.4. Quaternary Re- search Association, Cambridge, 45-83. Smart, P. L. and Richards, D. A., 2003. Paleoclimate determination from speleothems: progress, problems and pros- pects. Proceedings of the Conference on Climate changes: the karst record III, Montpellier, France, 11-14 May 2003. 27-29. Spötl, C. and Mattey, D., 2006. Stable isotope microsampling of speleothems for palaeoenvironmental studies: A comparison of microdrill, micromill and laser ablation techniques. Chemical Geology 235, 48-58. Stuiver, M., Reimer, P. J., Bard, E., Beck, J. W., Burr, G. S., Hughen, K., Kromer, B., McCormac, G., van der Plicht, J. and Spurk, M., 1998. INTCAL98 radiocarbon age calibration, 24,000-0 cal BP. Radiocarbon 40(3), 1041-83. Sundqvist, H. S., Holmgren, K. and Lauritzen, S. E. 2007. Sta- ble isotope variations in stalagmites from northwestern Swe- den document climate and environmental changes during the early Holocene. The Holocene 17(1). Tan, M., Baker, A., Genty, D., Smith, C., Esper, J. and Binggui, C., 2006. Applications of stalagmite laminae to paleoclimate reconstructions: Comparison with dendrochronology/climatology. Quaternary Science Reviews 25, 2103-2117. Thompson, P., Schwarcz, H. P. and Ford, D. C., 1974. Conti- nental Pleistocene climatic variations from speleothem age and isotopic data. Science 184, 893-894. Vollweiler, N., Scholz, D., Mühlinghaus, C., Mangini, A. and Spötl, C., 2006. A precisely dated climate record for the last 9 kyr from three high alpine stalagmites, Spannagel Cave, Austria. Geophysical Research Letters, Vol. 33, L20703, doi:10.1029/2006GL027662, 2006. Wigley, T. M. L., and Brown, M. C., 1976. The physics of caves. In: The Science of speleology. Ford, D. T., Cullingford, C. H. C., eds. Academic Press, London, 329-358. Younge, C. J., 1981. Fluid inclusions in speleothems as palaeoclimatic indicators. In: Proceedings of the 8th International Congress of Speleology 1: 301-304. Younge, C. J., Ford, D. C., Gray, J. and Schwarcz, H. P., 1985. Stable isotope studies of cave seepage water. Chemical Ge- ology 58, 97-105.